by
Francisco Alberto SANDOVAL NOREÑA
MANUSCRIPT-BASED THESIS PRESENTED TO ÉCOLE DE
TECHNOLOGIE SUPÉRIEURE IN PARTIAL FULFILLMENT FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY
Ph.D.
MONTREAL, JULY 03, 2019
ÉCOLE DE TECHNOLOGIE SUPÉRIEURE
UNIVERSITÉ DU QUÉBEC
BY THE FOLLOWING BOARD OF EXAMINERS
Pr. François Gagnon, Thesis Supervisor
Department of Electrical Engineering, École de technologie supérieure
Dr. Gwenael Poitau, Co-supervisor
Chief Technology Officer, Ultra Electronics
Pr. Ismail Ben Ayed, President of the Board of Examiners
Artificial Intelligence in Medical Imaging, École de technologie supérieure
Pr. Georges Kaddoum, Member of the jury
Department of Electrical Engineering, École de technologie supérieure
Pr. Tayeb A. Denidni, External Examiner
Énergie Matériaux Télécommunications Research Centre, Institut national de la recherche scientifique
THIS THESIS WAS PRESENTED AND DEFENDED
IN THE PRESENCE OF A BOARD OF EXAMINERS AND THE PUBLIC ON "DEFENSE DATE"
continuous support and help of my Ph.D. study and related research. From the beginning he has trusted me. Thanks for your guidance and especially for your friendship. I also want to thank my co-asvisor Dr. Gwenael Poitau, and through him Ultra Electronic, for the opportunity provided to work together, and I acknowledge Eric Kwati who helped me improve my English writing.
I am grateful to the Natural Sciences and Engineering Research Council of Canada, and the Ultra Electronics TCS through the Industrial Research Chair in High Performance Wireless Emergency and Tactical Communications for funding, in part, this work.
I want to thank my committee members, Professor Ismail Ben Ayed (President of the Board of Examiners), Professor Georges Kaddoum (Member of the jury), Professor Tayeb A. Denidni. (External Examiner) for serving as my committee members and I also want to thank you for your interesting comments and suggestions, thanks to you.
I thank the UTPL for the support provided, especially the colleagues of the Electronics and Telecommunications Section, and the Department of Computer Science and Electronics. Un profundo agradecimiento para Glenda, mi esposa, quien desde un principio ha apoyado mis sueños a pesar de los sacrificios que traen consigo, quien ha sufrido, se ha alegrado y me ha acompañado de muchas maneras durante este camino, con mucho amor y dedicación. Gracias de todo corazón.
Gratitud inmensa para mis padres Lubín y Fabiola, mis hermanos José y María Fabiola, y toda mi familia ya que siempre han creído en mí y me han brindado sus oraciones, su apoyo, su ánimo y alegría para superar los obstáculos y lograr las metas propuestas.
Agradezco también a todas las amigas y los amigos que de una u otra manera han apoyado este caminar y me han animado a seguir adelante.
I also place on record, my sense of gratitude to one and all, who directly or indirectly, have helped and supported me during my work in the ÉTS and my life in Canada, especially the friends with whom we have shared this process: Byron and family, Tuesman, Katty, Alex, Gabriel and family, Ruth, and Verónica.
RÉSUMÉ
Actuellement, les systèmes MIMO-OFDM constituent la base d’importants systèmes de com-munication sans fil tels que les réseaux commerciaux 4G et 5G, les comcom-munications tactiques et les communications interopérables pour la sécurité publique. Cependant, un inconvénient de la modulation OFDM est le haut rapport puissance crête à puissance moyenne (PAPR). Ce problème augmente lorsque le nombre d’antennes d’émission augmente.
Ce travail consiste à proposer une nouvelle technique de réduction de PAPR hybride pour les systèmes MIMO-OFDM de codage de bloc espace-temps (STBC). Cette technique combine les capacités de codage aux méthodes de réduction des PAPR tout en exploitant le nouveau degré de liberté offert par la présence de plusieurs chaînes de transmission (MIMO).
Dans la première partie, nous présentons une revue de littérature approfondie des techniques de réduction de PAPR pour les systèmes OFDM et MIMO-OFDM. Ces travaux ont permis de mettre au point une taxonomie de technique de réduction de PAPR, d’analyser les motivations de réduction de PAPR dans les systèmes de communication actuels tenant compte du gain de couverture, de comparer par simulation les caractéristiques de chaque catégorie et conclure par l’importance des techniques de réduction de PAPR hybride.
Dans la deuxième partie, nous étudions l’effet des codes de correction d’erreurs en aval (FEC), tels que les codes de bloc linéaires et les codes de convolution sur le PAPR du système OFDM codé (COFDM). Nous avons simulé et comparé la fonction de distribution cumulative com-plémentaire (FDCC) du PAPR et sa relation avec l’autocorrélation du signal COFDM avant de transiter par le bloc de transformée de Fourier rapide inverse (IFFT). Cela permet de con-clure sur les caractéristiques principales des codes qui génèrent des pics élevés dans le signal COFDM, et donc sur les paramètres optimaux afin de réduire le PAPR.
Enfin, nous proposons une nouvelle technique de réduction de PAPR hybride pour le sys-tème STBC MIMO-OFDM, dans laquelle le code de convolution est optimisé pour éviter la dégradation de PAPR, et combine les schémas successifs de rotation, d’inversion (SS-CARI), de composition modifiée itérative et de filtrage sous-optimal. La nouvelle méthode permet d’obtenir un gain net considérable pour le système, c’est-à-dire une réduction considérable du PAPR, un gain de taux d’erreur sur les bits (BER) par rapport au système MIMO-OFDM de base, une faible complexité et une empreinte spectrale réduite. La nouvelle technique hybride a été largement évaluée par simulation et la fonction de distribution cumulative complémentaire (CCDF), le BER et la densité spectrale de puissance (PSD) ont été comparés au signal STBC MIMO-OFDM d’origine.
Mots-clés: Compression, codes convolutifs, amplificateur de puissance élevée, entrées multiples sorties multiples (MIMO), multiplexage par répartition orthogonale de la fréquence (OFDM), rapport de puissance pic à moyenne (PAPR), codage de bloc espace-temps (STBC), communication tactique
ABSTRACT
Currently, multiple-input multiple-output orthogonal frequency division multiplexing (MIMO-OFDM) systems underlie crucial wireless communication systems such as commercial 4G and 5G networks, tactical communication, and interoperable Public Safety communications. How-ever, one drawback arising from OFDM modulation is its resulting high peak-to-average power ratio (PAPR). This problem increases with an increase in the number of transmit antennas. In this work, a new hybrid PAPR reduction technique is proposed for space-time block coding (STBC) MIMO-OFDM systems that combine the coding capabilities to PAPR reduction meth-ods, while leveraging the new degree of freedom provided by the presence of multiple transmit chairs (MIMO).
In the first part, we presented an extensive literature review of PAPR reduction techniques for OFDM and MIMO-OFDM systems. The work developed a PAPR reduction technique taxo-nomy, and analyzed the motivations for reducing the PAPR in current communication systems, emphasizing two important motivations such as power savings and coverage gain. In the tax onomy presented here, we include a new category, namely, hybrid techniques. Additionally, we drew a conclusion regarding the importance of hybrid PAPR reduction techniques.
In the second part, we studied the effect of forward error correction (FEC) codes on the PAPR for the coded OFDM (COFDM) system. We simulated and compared the CCDF of the PAPR and its relationship with the autocorrelation of the COFDM signal before the inverse fast Fourier transform (IFFT) block. This allows to conclude on the main characteristics of the codes that generate high peaks in the COFDM signal, and therefore, the optimal parameters in order to reduce PAPR. We emphasize our study in FEC codes as linear block codes, and convolutional codes.
Finally, we proposed a new hybrid PAPR reduction technique for an STBC MIMO-OFDM system, in which the convolutional code is optimized to avoid PAPR degradation, which also combines successive suboptimal cross-antenna rotation and inversion (SS-CARI) and iterative modified companding and filtering schemes. The new method permits to obtain a significant net gain for the system, i.e., considerable PAPR reduction, bit error rate (BER) gain as com-pared to the basic MIMO-OFDM system, low complexity, and reduced spectral splatter. The new hybrid technique was extensively evaluated by simulation, and the complementary cu-mulative distribution function (CCDF), the BER, and the power spectral density (PSD) were compared to the original STBC MIMO-OFDM signal.
Keywords: Compression, convolutional codes, high power amplifier, multiple-input multiple-output (MIMO), orthogonal frequency division multiplexing (OFDM), peak-to-average power ratio (PAPR), space-time block coding (STBC), tactical communication
INTRODUCTION . . . 1
CHAPTER 1 MIMO-OFDM PAPR REDUCTION TECHNIQUES: DESCRIPTION AND SYSTEMATIC REVIEW . . . 7
1.1 Introduction . . . 7
1.2 Background . . . 8
1.2.1 MIMO-OFDM system . . . 8
1.2.2 PAPR problem . . . 10
1.2.3 PAPR reduction techniques . . . 10
1.3 Research methodology . . . 12
1.3.1 Research questions . . . 13
1.3.2 Search strategy . . . 13
1.3.3 Study selection and quality assessment . . . 15
1.3.4 Data extraction . . . 17
1.3.5 Data analysis . . . 18
1.4 Results of systematic literature review . . . 18
1.4.1 Frequency of publication (RQ1) . . . 18
1.4.2 Download Database . . . 19
1.4.3 MIMO-OFDM system applied (RQ2) . . . 19
1.4.4 Proposed classification of PAPR reduction techniques (RQ3) . . . 22
1.4.5 Classification of MIMO-OFDM PAPR reduction techniques based on its adaptation to the MIMO structure (RQ4) . . . 23
1.4.6 Most used PAPR reduction techniques (RQ3) description . . . 25
1.4.6.1 Coding techniques . . . 25
1.4.6.2 Precoding techniques . . . 26
1.4.6.3 Selected mapping techniques . . . 26
1.4.6.4 Partial transmit sequence techniques . . . 27
1.4.6.5 Tone reservation techniques . . . 28
1.4.6.6 Clipping techniques . . . 29
1.4.6.7 Hybrid techniques . . . 30
1.5 Conclusion . . . 30
CHAPTER 2 HYBRID PEAK-TO-AVERAGE POWER RATIO REDUCTION TECH-NIQUES: REVIEW AND PERFORMANCE COMPARISON . . . 33
2.1 Introduction . . . 34
2.2 OFDM System Model and PAPR Problem . . . 36
2.2.1 The CCDF of the PAPR . . . 38
2.2.2 Net gain . . . 40
2.3 Motivation . . . 41
2.3.2 Coverage gain . . . 45
2.4 PAPR Reduction Techniques . . . 50
2.4.1 Coding Based Techniques . . . 51
2.4.1.1 Simple Odd Parity Code . . . 52
2.4.1.2 Modified Code Repetition . . . 53
2.4.1.3 Complement Block Coding . . . 54
2.4.1.4 Sub-block complementary coding . . . 54
2.4.1.5 Golay complementary sequences . . . 55
2.4.2 Multiple Signaling and Probabilistic Techniques . . . 55
2.4.2.1 Selected Mapping . . . 55
2.4.2.2 Partial Transmit Sequence (PTS) . . . 57
2.4.2.3 Interleaving . . . 58
2.4.2.4 DFT-Spreading Technique . . . 59
2.4.2.5 Tone Reservation . . . 60
2.4.2.6 Tone Injection . . . 61
2.4.2.7 Dummy Sequence Insertion . . . 61
2.4.3 Signal Distortion Techniques . . . 62
2.4.3.1 Amplitude Clipping . . . 62
2.4.3.2 Peak Windowing . . . 63
2.4.3.3 Companding . . . 63
2.4.4 Hybrid Techniques . . . 64
2.4.4.1 Partial Transmit Sequence Using Error-Correcting Code (PTS-ECC) . . . 65
2.4.4.2 Error Control Selected Mapping (EC-SLM) . . . 66
2.4.4.3 Error Control Selected Mapping with Clipping (EC-SLM-CP) . . . 66
2.5 Modified Code Repetition, Selected Mapping and Clipping (MCR-SLM-CP) . . . 68
2.5.1 Comparison of PAPR Reduction Techniques . . . 69
2.6 Conclusion . . . 72
CHAPTER 3 OPTIMIZING FORWARD ERROR CORRECTION CODES FOR COFDM WITH REDUCED PAPR . . . 75
3.1 Introduction . . . 76
3.2 Background . . . 78
3.2.1 Forward Error Correction . . . 78
3.2.1.1 Block Codes . . . 79
3.2.1.2 Convolutional Codes . . . 79
3.2.2 COFDM System Model . . . 81
3.2.3 PAPR Problem . . . 81
3.2.4 Net gain . . . 82
3.2.5 Distribution of the PARP for COFDM Signal . . . 83
3.3.1 Autocorrelation characteristics of uncoded OFDM signal . . . 85
3.3.2 Markov Chain model for autocorrelation of coded OFDM . . . 86
3.3.3 Upper bound on peak factor of coded OFDM signal analysis . . . 87
3.3.3.1 Linear block code . . . 87
3.3.3.2 Convolutional Code . . . 90
3.4 Analysis of PAPR Degradation in COFDM . . . 92
3.4.1 Linear block code: Repetition code . . . 92
3.4.2 Convolutional codes . . . 94
3.4.2.1 Code rate . . . 96
3.4.2.2 Code structure . . . 98
3.4.2.3 Maximum free distance Convolutional codes . . . .100
3.4.2.4 Constraint length . . . .102
3.5 Optimal Convolutional Code to Avoid an Increase in the PAPR Based on Net Gain . . . .104
3.5.1 Decoder complexity . . . .108
3.6 Conclusion . . . .109
CHAPTER 4 ON OPTIMIZING THE PAPR OF OFDM SIGNALS WITH CODING, COMPANDING, AND MIMO . . . .111
4.1 Introduction . . . .112
4.2 Background . . . .114
4.2.1 System Model . . . .114
4.2.2 Convolutional Code . . . .116
4.2.3 Successive Suboptimal Cross-Antenna Rotation and Inversion (SS-CARI) Scheme . . . .116
4.2.4 Modifiedμ-Law Companding . . . .118
4.3 Proposed Hybrid PAPR Reduction Technique . . . .119
4.4 Performance of Hybrid PAPR Reduction Technique . . . .121
4.5 Conclusion . . . .127
CONCLUSION AND RECOMMENDATIONS . . . .131
APPENDIX I SYSTEMATIC LITERATURE REVIEW RESULTS . . . .139
Table 1.1 Keywords and Synonyms . . . 14
Table 1.2 Searches in databases . . . 15
Table 1.3 Summary of search results . . . 15
Table 1.4 Number of studies per study selection . . . 16
Table 1.5 Quality Assessment: Answers . . . 17
Table 1.6 Quality Assessment Score . . . 17
Table 1.7 Data extraction form . . . 18
Table 2.1 A comparison of the PA efficiency with and without PAPR reduction of different PA classes . . . 45
Table 2.2 Commercial and tactical communications parameters comparison (Ozaet al., 2012) . . . 47
Table 2.3 PAPR Reduction Hybrid Techniques. . . 65
Table 2.4 Net gain of PAPR reduction techniques. . . 72
Table 3.1 Register Contents, Output Code Words and Elements of the Code Word Matrix for Convolutional Code(2,1,3). . . 92
Table 3.2 Λfor COFDM used in the linear block code examples . . . 95
Table 3.3 States, Register Contents, and Output Code Words for Convolutional code: (V =3;[5,7]),(V =3;[1,3,7,3]), and(V =3;[1,5,7,3,1,5,3,7]) . . . 97
Table 3.4 States, Register Contents, and Output Code Words for Convolutional codes: (V =3;[1,3,5,7]),(V =3;[5,5,7,7]),(V =3;[5,7,7,7]), and(V = 3;[7,7,7,7]) . . . 99
Table 3.5 Maximum Free Distance Codes (Proakis & Salehi, 2008) with code rate 1/2, 1/4, and 1/8, and the structure number (ς) for each code . . . .102
Table 3.7 Net gain for different maximum free distance convolutional code
with code rate equal to 1/2,1/4,and 1/8 . . . .105 Table 4.1 Rate 1/4 Maximum Free Distance Codes (Proakis & Salehi, 2008). . . .123
Figure 1.1 MMO-OFDM generic model . . . 9
Figure 1.2 Taxonomy for PAPR reduction techniques . . . 12
Figure 1.3 Final articles per year (2017, before of 25 July). . . 19
Figure 1.4 Articles per download database . . . 20
Figure 1.5 Purpose MIMO taxonomy Hampton (2013) . . . 21
Figure 1.6 Articles per MIMO system type. . . 22
Figure 1.7 Articles per proposed PAPR reduction technique type . . . 23
Figure 1.8 Articles per PAPR reduction technique . . . 24
Figure 1.9 Articles per technique approach . . . 25
Figure 1.10 SISO-OFDM SLM generic scheme Choet al.(2010a) . . . 27
Figure 1.11 SISO-OFDM PTS generic scheme, Choet al.(2010a) . . . 28
Figure 1.12 SISO-OFDM tone reservation generic scheme, Choet al.(2010a) . . . 29
Figure 1.13 SISO-OFDM clipping generic scheme, Choet al.(2010a) . . . 30
Figure 2.1 Block diagram of transmitter and receiver in an OFDM system . . . 36
Figure 2.2 Theoretical CCDFs of OFDM signals with different subcarriers . . . 39
Figure 2.3 Time domain OFDM signals withK =4 for real, imaginary parts and the sum |x(t)|, when the modulation is QPSK (Cho et al., 2010a) . . . 41
Figure 2.4 Effects of nonlinear PA on (a) signal spectrum and (b) signal constellation (Ramezani, 2007) . . . 42
Figure 2.5 Input power versus output power characteristics and efficiency curves for a solid state power amplifier (SSPA) . . . 43
Figure 2.6 Range extension (left) and Coverage area (right) as a function of transmit power gaingP (Khan, 2009) . . . 49
Figure 2.7 PAPR reduction techniques. . . 52
Figure 2.8 PAPR of a four subcarrier signal for all possible data wordsdn. . . 53
Figure 2.9 Examples of Coding-based techniques . . . 54
Figure 2.10 Block diagram of selected mapping technique for PAPR reduction . . . 56
Figure 2.11 Block diagram of PTS technique for PAPR reduction (Cho et al., 2010a) . . . 58
Figure 2.12 Block diagram for single carrier-FDMA (SC-FDMA) technique for PAPR reduction . . . 59
Figure 2.13 Block diagram of tone reservation technique for PAPR reduction (Choet al., 2010a) . . . 60
Figure 2.14 Block diagram of tone injection technique for PAPR reduction (Choet al., 2010a) . . . 61
Figure 2.15 Block diagram of peak windowing technique for PAPR reduction (Rahmatallah & Mohan, 2013) . . . 64
Figure 2.16 Block diagram of an EC-SLM transmitter and receiver (Xin & Fair, 2004) . . . 67
Figure 2.17 Block diagram of EC-SLM-CP technique (Carson & Gulliver, 2002) . . . 67
Figure 2.18 Block diagram of MCR-SLM-CP hybrid technique . . . 69
Figure 2.19 Comparisons of CCDF in OFDM-BPSK system for PAPR reduction techniques with Ns = 3e+5 for conventional OFDM, CP 70%, CP 50%, and MCR (R =1/4), and Ns = 1e+5 for SLM (U =4), MCR+SLM (U = 4)+CP 70%, and MCR+SLM (U =8)+CP 50% . . . 70
Figure 2.20 Comparisons of BER in OFDM-BPSK system for PAPR reduction techniques . . . 71
Figure 3.1 Block diagram of transmitter and receiver in a coded OFDM system . . . 82
Figure 3.2 (a) Autocorrelation of repetition code with η = 4, (b) Autocorrelation of repetition code withη =8 . . . 89
Figure 3.3 (a) (2,1,3) convolutional encoder, (b) state diagram for (2,1,3)
convolutional encoder (Sklar, 2001) . . . 91
Figure 3.4 COFDM system with repetition code with code rate 1/2,1/4,and 1/8,N=512 subcarriers, andNs=105OFDM symbols simulated. . . 93
Figure 3.5 COFDM system with repetition code, repetition code plus interleaving and MCR plus interleaving with code rate 1/4, N = 512 subcarriers, andNs=105OFDM symbols simulated. . . 94
Figure 3.6 COFDM system with repetition code, repetition code plus interleaving and MCR plus interleaving with code rate R=1/8, N=512 subcarriers, andNs=105OFDM symbols simulated. . . 95
Figure 3.7 COFDM-QPSK system for different convolutional code rate with N=256 subcarriers, andNs=105OFDM symbols simulated. . . 98
Figure 3.8 COFDM-QPSK system for different convolutional code structures with code rateR=1/4,N=256 subcarriers, andNs=105OFDM symbols simulated. . . .100
Figure 3.9 COFDM-QPSK system for different maximum free distance convolutional codes with N =256 subcarriers, constraint length V between 3 and 8,Ns=105OFDM symbols simulated, and code ratesR=1/2,R=1/4, andR=1/8. . . .102
Figure 3.10 Autocorrelation for COFDM-QPSK system with N = 256 subcarriers, Ns = 105 OFDM symbols simulated, different maximum free distance convolutional codes, and code rates R= 1/2,R=1/4, andR=1/8. . . .103
Figure 3.11 Algorithm to calculate the optimal code to avoid an increase in the PAPR based on the net gain . . . .106
Figure 3.12 (a) Net gain with α1 = α2 = 0.5 for different maximum free distance convolutional codes with code rate 1/2, 1/4, and 1/8, (b)Y1 versusY2for different constraint length (V) with code rate 1/2,1/4, and 1/8 . . . .107
Figure 4.1 SS-CARI algorithm. . . .117
Figure 4.2 Block diagram of the CSC technique. . . .120
Figure 4.3 Frequency domain filtering based on (Armstrong, 2001). . . .121
Figure 4.5 Comparison of CCDF in MIMO-OFDM system for different maximum free distance convolutional codes with R =1/4. The
reference is the STBC MIMO-OFDM system without coding. . . .122 Figure 4.6 Comparison of CCDF in MIMO-OFDM system with SS-CARI
technique with different numbers of subblocks. The reference is
the STBC MIMO-OFDM system without coding. . . .123 Figure 4.7 BER vs G with modified companding transforms (QPSK, L=4,
N=128, CCDF=10−3, and SNR=12 dB). . . .124 Figure 4.8 CCDF of PAPR curves for CSC hybrid PAPR reduction technique
in STBC MIMO-OFDM system withR=1/4,V =4,M=16, and two versions: μ=10, PR=1.2 andκ=2.7 (case 1), andμ=255, PR=2 andκ=15.7 (case 2). The reference is the STBC
MIMO-OFDM system without coding. . . .125 Figure 4.9 BER performance for CSC hybrid PAPR reduction technique in
STBC MIMO-OFDM system withR=1/4, V =4, M=16, and two versions: μ=10, PR=1.2 andκ=2.7 (case 1), andμ=255, PR=2 andκ=15.7 (case 2). The reference is the STBC
MIMO-OFDM system without coding. . . .126 Figure 4.10 PSD curves for CSC hybrid PAPR reduction technique in STBC
MIMO-OFDM system with R= 1/4, V =4, M =16, and two versions: μ =10, PR=1.2 and κ =2.7 (case 1), and μ =255, PR=2 andκ=15.7 (case 2). The reference is the STBC
MIMO-OFDM system without coding. . . .127 Figure 4.11 CCDF of PAPR curves for CSC hybrid PAPR reduction technique
with iterative MuCT and filter in STBC MIMO-OFDM system with R=1/4, V =4, M =16, μ =10, PR =1.2, and κ =2.7 (case 1). The reference is the STBC MIMO-OFDM system without
coding. . . .128 Figure 4.12 BER performance for CSC hybrid PAPR reduction technique with
iterative MuCT and filter in STBC MIMO-OFDM system with R=1/4,V =4,M=16, μ =10, PR=1.2, andκ =2.7 (case 1).
The reference is the STBC MIMO-OFDM system without coding. . . .129 Figure 4.13 PSD curves for CSC hybrid PAPR reduction technique with
iterative MuCT and filter in STBC MIMO-OFDM system with R=1/4,V =4,M=16, μ =10, PR=1.2, andκ =2.7 (case 1).
AWGN Additive white Gaussian noise BER Bit error rate
BLAST Bell Laboratories layered space-time
BS Base station
CARI Cross-antenna rotation and inversion CBC Complement block coding
CDF Cumulative distribution function
CCDF Complementary cumulative distribution function
C+MSP Coding plus Multiple Signaling and Probabilistic techniques
C+MSP+SD Coding plus Multiple Signaling and Probabilistic techniques plus Signal Dis-tortion techniques
CP Cyclic prefix
COFDM Coded orthogonal frequency division multiplexing C+SD Coding plus Signal Distortion techniques
DAB Digital Audio Broadcasting DAC Digital to analog converter
D-BLAST Diagonal Bell Laboratories layered space-time dSLM Directed selected mapping
DVB-S Digital Video Broadcasting-Satellite DVB-T Digital Video Broadcasting-Terrestrial
eNB Node-B
ECC Error-correcting codes
EC-SLM Error control selected mapping
EC-SLM-CP Error control selected mapping with clipping FEC Forward error correction
HARQ Hybrid automatic repeat request
H-BLAST Horizontal Bell Laboratories layered space-time HPA High power amplifier
IBO Input back-off
IFFT Inverse fast Fourier transform iSLM Individual selected mapping
ITU International Telecommunication Union JTRS Joint Tactical Radio System
LDPC Low-density parity-check codes LST Less significant bit
LTE Long-Term Evolution MCR Modified code repetition
MGSTC Multi-group space time coding MIMO Multiple-input multiple-output MME/GT Mobility management entity/gateway MSP Multiple signaling and probabilistic
MSP+SD Multiple signaling and probabilistic plus Signal Distortion techniques MuCT Modifiedμ-law compander transform
MU-MIMO Multiple-user multiple input multiple output NCT Nonlinear companding transform
NOSTBC Non-orthogonal space-time block codes OFDM Orthogonal frequency division multiplexing
OFDMA Orthogonal frequency division multiplexing access oSLM Ordinary selected mapping
OSTBC Orthogonal space-time block codes
PA Power amplifier
PAPR Peak-to-average power ratio
PMEPR Peak-to-mean envelope power ratio PSD Power spectral density
PR Peak ratio
PRT Peak reduction tones P/S Parallel-to-serial
PTS Partial transmit sequence
PTS-ECC Partial transmit sequence Using Error-Correcting Code QPSK Quadrature phase-shift keying
QAM Quadrature amplitude modulation QoS Quality of service
RA Repeat accumulate codes RC Repetition codes
RM Reed-Muller
SBCC Sub-block complementary coding
SC-FDMA Single Carrier Frequency Division Multiple Access SD Signal distortion
SISO Single-input single-output SLM Selected mapping
S/P Serial-to-parallel SRW Soldier radio waveform
SS-CARI Successive suboptimal cross-antenna rotation and inversion SSPA Solid state power amplifier
SOPC Simple odd parity code STBC Space-time block codes STC Space-time coding
STTC Space-time trellis codes
SU-MIMO single-user multiple input multiple output TI Tone injection
TR Tone reservation
TSTC Threaded space-time coding uCT μ-law compander transform
UE User equipment
UNW Universal Networking Waveform
V-BLAST Vertical Bell Laboratories layered space-time WEA Wavelet entropy algorithm
WNW Wideband Network Waveform
WiMAX Worldwide Interoperability for Microwave Access 4G 4th generation
ci Code word vector
E[·] Expected value
Es Symbol energy
Esc Symbol energy of the companded signal
f Frequency
G Generation matrix GR Receiver antenna gain
GT Transmitter antenna gain
hi j Channel from theith transmit antenna to the jth receive antenna
Hi j[k] Channel frequency response of the(i,j)th channel Ik k×kIdentity matrix
j Imaginary unit
K Number of subcarriers (OFDM) (Chapter 2) K Number of states (Chapter 3)
K Normalization constant (Chapter 4) L Maximum length of channel
N Number of subcarriers (OFDM) Ncp Length of the cyclic prefix
NR Number of receive antennas
Nt Number of transmit antennas
P Parity matrix
PR Received power
PT Radiated power
PAPR(·) peak-to-average power ratio PLFS Path loss in free space
PR Peak ratio
Pr(·) Probability distribution function
R Code rate
R(k) State correlation matrix ts sampling period
U Number of phase sequences (SLM)
V Number of subsequences (PTS) (Chapter 2) V Peak amplitude of compressor (Chapter 4)
xi(n) Discrete-time baseband OFDM signal from theith transmit antenna
x(n) Discrete-time baseband OFDM signal x(t) continuous-time baseband OFDM signal X(k) Data symbol (OFDM)
{Xi[k]}N−k=01 Frequency-domain signal from theith transmit antenna
xpeak Peak of the actual signal
Yj[k] Received frequency-domain signal at the jth receive antenna ΔA Gain in coverage area
ΔR Incremental range extension η Power amplifier efficiency
π Number pi
λ Wavelength
α Path loss exponent
Π
ΠΠ Transition probability matrix
dB DeciBels
cause the network density in it is lower than in commercial communications and thus increased range is key to maximize the coverage.
Multiple tactical waveforms, such as the Universal Networking Waveform (UNW) and the Wideband Network Waveform (WNW), are based on orthogonal frequency division multiplex-ing (OFDM) for inherent mobility robustness. OFDM is present in all 4G wireless communi-cation systems, including the IEEE 802.16 Worldwide Interoperability for Microwave Access (WiMAX) and Long Term Evolution (LTE) standards. Furthermore, OFDM is a popular mo-dulation for other communication systems; for example, it is included in IEEE 802.11 a/g/n/ac wireless LANs, Digital Audio Broadcasting (DAB), Digital Video Broadcasting-Terrestrial (DVB-T), and Digital Video Broadcasting by Satellite (DVB-S). Additionally, OFDM is part of the 5G waveforms proposal.
Multi-carrier modulation OFDM has multiples advantages, such as high spectral efficiency, high data rate transmission over a multipath fading channel, simple implementation by the Fast Fourier transformation (FFT), and low receiver complexity (Yang, 2005). However, a critical drawback of OFDM is the high peak-to-average power ratio (PAPR) it produces, which can impact the performance of the non-linear elements in the system, such as the high power amplifier (HPA) and the digital-to-analog converter (DAC). Besides, the high picks in the multi-carrier signal can indirectly influence the system’s range and coverage or power consumption (Sandovalet al., 2017).
The problem of PAPR in OFDM modulation has been widely analyzed in the literature, as presented in Chapters 1 and 2, where PAPR reduction techniques are organized into four ca-tegories, namely, Coding, Multiple Signaling and Probabilistic (MSP), Signal Distortion (SD), and Hybrid. In Coding schemes, codewords that minimize the PAPR are selected, while
mul-tiple permutation of the multi-carrier signal is generated in Mulmul-tiple Signal techniques, and the signal with the minimum PAPR is chosen for transmission. For their part, the probabilistic methods modify different parameters in the multi-carrier signal and optimize them to minimize the PAPR. Finally, the hybrid schemes combine two or more techniques for PAPR reduction in a bid to take advantage of different techniques that can result in greater PAPR reduction, better performance, and stronger control of parameters such as computational complexity, additional transmission power requirements, and data rate drops.
Combining OFDM with multiple-input multiple-output (MIMO) wireless communication sys-tems results in MIMO-OFDM, one of the most used techniques in current syssys-tems, and the most promising. On the other hand, MIMO-OFDM presents the problem of a high PAPR, re-sults in a higher PAPR than OFDM alone since the PAPR increases as the number of transmit antennas increases as well.
Research Objectives
The main contribution of this thesis is to combine coding capabilities with PAPR reduction methods while leveraging the new degree of freedom provided by the presence of multiple transmit chains (MIMO). The research objectives include:
1. Analyze the main performance improvements for PAPR reduction in current communica-tions systems with an emphasis on tactical communicacommunica-tions systems.
2. Explore OFDM PAPR reduction techniques proposed in the literature, their classification, their characteristics and their possibilities of being combined with hybrid techniques. 3. Study, model, and compare the PAPR reduction capabilities of forward error correction
codes.
4. Examine the proposed PAPR reduction techniques for MIMO-OFDM systems and check which of the systems have better characteristics that can be combined with coding tech-niques.
5. Propose a novel hybrid peak-to-average power ratio reduction scheme that combines a code-based technique with signal distortion or multiple signaling and a probabilistic method in a tactical MIMO-OFDM system to reduce the PAPR and to achieve BER reduction with-out increasing the saturation point of a power amplifier.
Methodology
This research begins with an extensive review of the literature on OFDM systems and PAPR reduction techniques. The work develops a systematic approach for PAPR reduction under different propagation, topology or traffic conditions, and presents a detailed analysis of the motivations for reducing the PAPR in current communication systems, emphasizing the re-sulting coverage gain. The thesis summarizes the recent literature on hybrid PAPR reduction techniques, compares the important parameters it incorporates, and concludes on its usability in current commercial, public safety, and tactical communications systems.
An OFDM communication system with different PAPR reduction techniques, including mul-tiple signals and probabilistic (MSP) and signal distortion (SD) methods, was simulated over an additive white Gaussian noise (AWGN) channel. The results compared the net gain perfor-mance based on the complementary cumulative distribution function (CCDF) and the bit error rate (BER). Additionally, other parameters were taken into account for the final analysis as the need for side information, the technique computational complexity, the in-band or out-of-band radiation, or increased power requirements at the transmitter.
The effects of forward error correction (FEC) on the PAPR for the coded orthogonal frequency division multiplexing (COFDM) system, are explored. We conducted an analysis comparing the CCDF of PAPR to the autocorrelation of the COFDM signal before the inverse fast Fourier transform (IFFT) block. This allowed us to deduce the principal coding characteristics that
generate the peak factor after the IFFT block, following which we could then choose the pa-rameters of the coded structure in order to reduce the peak power.
A space-time block coding (STBC) MIMO-OFDM system over a Rayleigh channel was simu-lated based on commercial communication parameters. The most outstanding PAPR reduction techniques available in the literature were compared, and a selection of the best methods to combine in the new hybrid technique was made based on the work objectives.
We proposed a new hybrid PAPR reduction technique for an STBC MIMO-OFDM system, in which the convolutional code is optimized to avoid PAPR degradation, and combined the SS-CARI and an iteratively modified companding and filtering methods. The new technique aimed to achieve a significant net gain for the system, i.e., considerable PAPR reduction, BER gain as compared to the basic MIMO-OFDM system, low complexity, and reduced spectral splatter. The new hybrid technique was extensively evaluated, and the CCDF, the BER, and the PSD were compared with the original STBC MIMO-OFDM signal.
Thesis Organization
This thesis is composed of four themed chapters, set out as follows:
The first chapter presents a systematic literature review of the PAPR reduction techniques avail-able to MIMO-OFDM systems aiming to classify, analyze, and compare the documented meth-ods. First, the MIMO-OFDM system model and the PAPR formulation are introduced. We then go on to describe the systematic literature review methodology, present and analyze the results, and finally, present a conclusion on the literature review. The study classifies the techniques ac-cording to the following criteria: the MIMO system, the strategy applied in reducing the PAPR in the OFDM signal, and whether or not the method is an extension of a proposed scheme for OFDM or takes advantage of the MIMO structure in designing the new technique.
Chapter 2 presents our first paper, “Hybrid Peak-to-Average Power Ratio Reduction Tech-niques: Review and Performance Comparison”. This paper, which was published in IEEE Access, describes the PAPR problem and summarizes the best known OFDM signal reduction techniques. It analyzes the main motivations for reducing the PAPR in current communication systems, and then highlights two of them: power savings and coverage gain. Additionally, we present a completed taxonomy for PAPR reduction techniques broken down into four catego-ries, namely, coding, signal distortion, multiple signaling and probabilistic, and hybrid tech-niques. To conclude, the paper compares one scheme under each category in an OFDM binary phase shift keying (BPSK) modulation system over an AWGN channel, evaluates the CCDF of the PAPR and the BER performance, and summarizes the results based on the introduction of the net gain concept.
Chapter 3 presents the second paper,Optimizing Forward Error Correction Codes for COFDM with Reduced PAPR. It has been submitted to IEEE Transactions on Communications. This work studies the impact of FEC on the PAPR for the COFDM system. An analysis is done based on a study of the distribution of the PAPR for the COFDM signal, and the relation bet-ween the autocorrelation before the IFFT block in an uncoded and in a coded OFDM system, as well as the maximum PAPR of the COFDM. For this, a Markov Chain model for auto-correlation of the coded OFDM is suggested and related with the upper bound of the peak factor of the coded OFDM signal. The theory is tested by simulation over linear block and convolutional codes. In the case of convolutional codes, four parameters that can impact the PAPR degradation and the BER performance were studied in detail: the code rate, the code structure, the maximum free distance, and the constraint length. The results in this paper will contribute to correctly selecting the codes to be used in conjunction with an OFDM system to avoid increasing the PAPR.
The third paper,On Optimizing the PAPR of OFDM Signals with Coding, Companding, and MIMO, published in IEEE Access, is presented in Chapter 4. We proposed a new hybrid PAPR reduction method for the STBC MIMO-OFDM system, in which the convolutional code is optimized to avoid PAPR degradation, and combine the SS-CARI and the iteratively modified companding and filtering schemes. The new hybrid PAPR reduction technique is evaluated by simulation in STBC MIMO-OFDM QPSK modulation over the Rayleigh channel, and the main results and contribution are presented at the end of the chapter.
The final chapter brings together the different strands of the thesis, summarizes the main find-ings of this work, and identifies areas for further research.
List of contributions
The majority of these works have been published or submitted to international journals. They are listed as follows:
Journal papers
• “Hybrid Peak-to-Average Power Ratio Reduction Techniques: Review and Performance Comparison”in IEEE Access (published) (Sandovalet al., 2017);
• “Optimizing Forward Error Correction Codes for COFDM with Reduced PAPR”in IEEE Transactions on Communications (published) (Sandovalet al., 2019a);
• “On Optimizing the PAPR of OFDM Signals with Coding, Companding, and MIMO” in IEEE Access (published) (Sandovalet al., 2019b).
SYSTEMATIC REVIEW
1.1 Introduction
The use of multiple antennas at both the transmitter and the receiver is a standard method for improving the performance and increasing the capacity of wireless communications systems. A multiple-input multiple-output (MIMO) system in a wireless network provides, spatial di-versity and array and multiplexing gains (Gesbert et al., 2003). When a MIMO is used, the system capacity can be improved as compared to a single-input single-output (SISO) system with flat Rayleigh fading or narrowband channels (Foschini, 1996). However, when MIMO is used in wideband channels, the intersymbol interference (ISI) problem is compounded that is solved by using orthogonal frequency division multiplexing (OFDM) which is combined with MIMO to improve the capacity and achieve ISI mitigation.
One of the main implementation drawbacks of OFDM is the high peak-to-average power ratio (PAPR) it entails, in addition to which it also affects the MIMO-OFDM system. Currently, numerous works propose techniques to reduce the PAPR in MIMO-OFDM. The first ideas in this regard were in the form extensions of existing techniques for OFDM systems. Then, new proposals emerged aimed at taking advantage of the MIMO architecture. However, the sheer number of proposed techniques makes it difficult to compare, classify and discern the choice of PAPR reduction schemes for MIMO-OFDM systems. Furthermore, it should be recalled that there may be various MIMO applications and ways in which it can be implemented, and so appropriate technique for a specific situation must be determined.
Through a systematic literature review, this chapter reports the state of the art on the PAPR reduction techniques available to the MIMO-OFDM system. More than 100 papers were re-viewed, with the following being the main findings: more and more techniques are being proposed for reduce peaks in MIMO systems, and the favorite approach consists in taking
ad-vantage of the structure of the systems based on multiple signalling and probabilistic schemes such as selected mapping, partial transmit sequence, precoding, and more recently, hybrid tech-niques. Additionally, most of the schemes found in the literature are directed at generic and spatially diverse MIMO systems, although there are some contributions aimed at other types of implementations, such as spatial multiplexing, multi-user MIMO or massive MIMO. The chapter is organized as follows. Section 1.2 begins by laying out the theoretical dimensions of the research and looks at how the PAPR problem affects the MIMO-OFDM system, and describes a simplified model. The third section is concerned with the systematic literature review methodology used for this study. The results are presented and analyzed in Section 1.4, and the final section concludes the literature review.
1.2 Background
1.2.1 MIMO-OFDM system
In general, OFDM modulation and the MIMO system allow easy integration and increase of the spectral efficiency. Let us consider a MIMO-OFDM system with Nt transmit antennas,
Nr receive antennas, and N subcarriers. At each transmit antenna, the conventional OFDM
modulator is employed, i.e., to generate the OFDM symbol, the input signal is serial-to-parallel (S/P) converter followed by an inverse fast Fourier transform (IFFT). Each OFDM symbol is parallel-to-serial (P/S) converted and a cyclic prefix (CP) of lengthNcpis added. As expected,
the OFDM demodulator structure is used at each receiver antenna, as can be seen in Fig. 1.1. Let us assume that the discrete-time baseband equivalent channel between each transmit-to-receive antenna link has a frequency impulse response of maximum length L, and is quasi-static. Hence, the channel from theith transmit antenna to the jth receive antenna is:
hi j =
hi j[0],hi j[1],···,hi j[L−1] T.
S/P Add CP IFFT P/S Transmier Receiver S/P Add CP IFFT P/S Remove CP S/P FFT P/S Remove CP S/P FFT P/S
Figure 1.1 MMO-OFDM generic model
If we considerNIFFT points from theith transmit antenna,{Xi[k]}N−k=01represents the
frequency-domain signal, wherek is the frequency index, and the discrete-time baseband OFDM signal xi[n]after applying IFFT is given by:
xi[n] = 1 √ N N−1
∑
k=0 Xi[k]ej2π kn N, n=0,1,...,N−1, (1.2)wherendenotes the discrete-time index, and j is the imaginary unit.
Let us assume perfect time synchronization and that there is no inter-symbol interference (ISI) between OFDM symbols, i.e.,Ncp≥L. The received frequency-domain signal at the jth receive
antenna can be expressed as:
Yj[k] = Nt
∑
i=1
Hi j[k]Xi[k] +Wj, (1.3)
whereHi j[k]is the channel frequency response of the(i,j)th channel, and is equal to:
Hi j[k] = L−1
∑
l=0 hi j[l]ej2π kl N, (1.4)1.2.2 PAPR problem
Large peaks can be present in the instantaneous output of an OFDM signal at each transmit antenna, known as a PAPR. The PAPR of a discrete-time baseband OFDM signal is defined as the ratio between the maximum instantaneous power and its average power (Jiang & Wu, 2008), and from theith transmit antenna, it is:
PAPR(xi[n]) max 0≤n≤N−1|xi[n]| 2 1 N N−1
∑
n=0 |xi[n]|2 . (1.5)In the MIMO-OFDM system, the PAPR is defined as the maximum of all Nt PAPR values
evaluated in each MIMO path (Manassehet al., 2012), that is:
PAPRMIMO= max 1≤i≤Nt
PAPR(xi[n]). (1.6)
In the literature, the complementary cumulative distribution function (CCDF) is an important metric for comparing the performance of different systems related to the PAPR. It should be recalled that the PAPR is a random variable, and so the CCDF is the probability that the PAPR of the MIMO-OFDM signal exceeds a given thresholdγ (also represented by PAPR0 in this thesis), i.e.:
CCDF=Pr(PAPR>γ). (1.7)
1.2.3 PAPR reduction techniques
The best alternative for reducing the high PAPR in the MIMO-OFDM system is to try to de-crease the wide variations in the OFDM signal before tackling the nonlinear devices. Section 2.3 explain and demonstrates how to do this. There are many techniques proposed in the li-terature for reduction the PAPR in OFDM system, and some of these techniques have been adapted for use in MIMO-OFDM systems. The PAPR reduction techniques for OFDM sys-tems can be classified into four categories, according to our taxonomy (Sandovalet al., 2017),
namely: Coding (C), Multiple Signaling & Probabilistic (MSP), Signal Distortion (SD), and Hybrid (H).
For the PAPR Coding reduction technique, the idea is to select the code words that minimize the PAPR. Sandoval et al.(2017) classified the coding technique based on the coded scheme used, including block coding, convolutional codes, and concatenate coding schemes. Exam-ples of Coding methods are: Simple Odd Parity code (SOPC) (Wilkinson & Jones, 1995), Modified Code Repetition (MCR) (Ngajikinet al., 2003), Complement Block Coding (CBC) (Jiang & Zhu, 2005), Sub-block Complementary Coding (SBCC) (Jiang & Zhu, 2004b), and Golay Complementary Sequences (Wilkinson & Jones, 1995; Jianget al., 2004b).
Multiple Signaling techniques create multiple permutations of the signal and calculate the PAPR to select the signal with the minimum value for transmission. Meanwhile, Probabilis-tic techniques reduce the high peaks by changing and optimizing different parameters of the OFDM signal. The best-known Probabilistic techniques include: Selected Mapping (SLM) (Bäumlet al., 1996; Muller & Huber, 1997b), Partial Transmit Sequence (PTS) (Muller & Hu-ber, 1997b), Interleaving (Jayalath & Tellambura, 2000), DFT-Spreading or Single Carrier-FDMA (SD-Carrier-FDMA), Tone Reservation (TR) (Tellado-Mourelo, 1999), Tone Injection (TI) (Tellado-Mourelo, 1999), and Dummy Sequence Insertion (Ryuet al., 2004).
On the other hand, in Signal Distortion techniques, the signal is distorted before the power amplifier, in order to reduce the PAPR. For instance, Amplitude Clipping (O’Neill & Lopes, 1995), Peak Windowing (Nee & de Wild, 1998), Companding (Wanget al., 1999a), are meth-ods which distort the signal.
Hybrid methods combine two or more PAPR reduction schemes. Examples of such techniques are: Partial Transmit Sequence using Error-Correction Code (PTS-ECC) (Ghassemi & Gul-liver, 2010), Error Control Selected Mapping (EC-SLM) (Xin & Fair, 2004), and Error Control Selected Mapping with Clipping (EC-SLM-CP) (Carson & Gulliver, 2002).
A summary of the taxonomy for PAPR reduction techniques is shown in Fig. 1.2. Most of these PAPR reduction techniques for OFDM systems are presented in detail in Chapter 2, Section 2.4. The techniques used in MIMO-OFDM systems are discussed in Subsections 1.4.4, 1.4.5, and 1.4.6 of this chapter.
PAPR Reducon Techniques
Constrained constellaon shaping Tone reservaon (TR)
Tone injecon (TI) Dummy sequence inseron
Mulple Signaling
& Probabilisc (MSP)
Pre-distoron Envelope scaling Peak reducon carrier Random phase update Interleaving
Paral transmit sequence (PTS)
Selected mapping (SLM)
Constellaon shaping Acve constellaon extension Pre-coding or pulse shaping DFT spreading
Coding
Block coding
Convoluonal coding
Concatenate coding
Repeon codes
BCH codes
(Bose-Chaudhuri-Hocquenghem)
Tradion, Viterbi decoding Golay codes
Reed-Solomon codes Hamming codes
LDPC codes
(Low Density Parity Check)
Turbo codes
Reed-Solomon codes/ Viterbi algorithm
Signal Distoron
(SD) Hybrid
Clipping and ltering Peak windowing Companding Peak cancellaon Coding + SD MSP + SD Coding + MSP + SD Coding + MSP
Figure 1.2 Taxonomy for PAPR reduction techniques
1.3 Research methodology
The systematic literature review is a strict method that facilitates the identification, extraction, classification, and synthesis of available information in the literature regarding a specific re-search topic (Britto & Usman, 2015; Dyba et al., 2007). The systematic literature review is selected in this work to specify how research methodology and its implementation is based on
the guide proposed by Petersen et al. (Petersenet al., 2015). The process used to conduct this systematic literature review is detailed below.
1.3.1 Research questions
The objective of this mapping study is to understand the state of the art on PAPR reduction techniques in a MIMO-OFDM systems and to determine the most commonly suggested of these techniques in the literature, as well as their performance. Consequently, we propose the following research questions:
• Question 1 (RQ1): How many studies were published over the years?
• Question 2 (RQ2): What kind of MIMO system is analyzed?
• Question 3 (RQ3): Which technique is suggesting for reducing the PAPR in a MIMO-OFDM system?
• Question 4 (RQ4): Is the suggested technique adapted from OFDM or takes advantage of the structure of MIMO?
1.3.2 Search strategy
The search is based on the strategy suggested by Wohlinet al.(2012), which includes PICOC (population, intervention, comparison, outcomes, and context). These help determine key-words and formulate search strings according to the research questions (Petersenet al., 2015).
• Population: Solutions that implement a MIMO-OFDM PAPR reduction technique.
• Intervention: We do not have an explicit intervention to be investigated.
• Comparison: In this study, we compare different methods used to reduce the PAPR in the MIMO-OFDM system, in terms of the CCDF of PAPR, BER and other performance measures available.
• Outcomes: Improve the CCDF of PAPR in a MIMO-OFDM system.
• Context: Commercial, public safety, and tactical communication systems.
The keywords are extracted from the population, comparison, outcomes, and context. After determining all relevant keywords, we relate them with synonyms. The result of this work can
be seen in Table 1.1, where we list all the keywords, synonyms and its relation with PICOC elements.
Table 1.1 Keywords and Synonyms
Keyword Synonyms Related to
BER Bit error rate Comparison
CCDF Complementary cumulative distribution function Comparison, Outcomes
MIMO-OFDM Multiple antennas OFDM Population
PARP
PAR
Population Peak-to-average power ratio
Peak-to-mean envelope power ratio PMEPR
Reduction Population
Technique Method Population
Scheme
Based on the research questions, the PICOC criteria and the keywords found, we obtained the following search strings:
("MIMO-OFDM"OR"multiple antennas OFDM")AND("PAPR"OR"Peak-to-average power ratio"OR"PAR"OR"PMEPR"OR"peak-to-mean envelope power ratio") AND
("reduction "OR"technique"OR"method"OR"scheme")
The primary sources were selected based on the recommendation made by Dybaet al.(2007): IEEE Xplore1, ISI Web of Science2, Scopus3, and Inspec/Compendex4(Engineering village). These sources cover most of the important electrical engineering databases, such as IEEE, Springer, Elsevier, and the Wiley Online Library.
The search string was optimized for each database. The values used in each case are presented in Table 1.2. The search process was limited to the English language. In addition, in the case
1 http://ieeexplore.ieee.org 2 http://www.isiknowledge.com 3 http://www.scopus.com
of the Scopus source, the initial search gave a large number of documents, and so a new search was carried out, limiting the keywords to just the title and the abstract. The search summary is presented in Table 1.3.
Table 1.2 Searches in databases
Database Search
IEEE Digital Library
(MIMO-OFDM OR multiple antennas OFDM) AND (PAPR OR Peak-to-average power
ratioORPARORPMEPRORpeak-to-mean envelope power ratio)AND(reductionOR
techniqueORmethodORscheme) ISI Web of Science
(TS=(("MIMO-OFDM" OR "multiple antennas OFDM") AND ("PAPR" OR "Peak-to-average power ratio"OR"PAR"OR"PMEPR"OR"peak-to-mean envelope power ratio")
AND("reduction "OR"technique"OR"method"OR"scheme")))ANDIdioma: (English)
Scopus
TITLE-ABS-KEY ( ( "MIMO-OFDM" OR "multiple
anten-nas OFDM" ) AND ( "PAPR" OR "Peak-to-average power
ra-tio"OR"PAR"OR"PMEPR"OR"peak-to-mean envelope power ratio" )AND( "reduction "OR"technique"OR"method"OR"scheme" ) )AND(LIMIT-TO( LANGUAGE , "En-glish" ) )
El Compendex
(((("MIMO-OFDM"OR"multiple antennas OFDM")AND("PAPR"OR"Peak-to-average power ratio" OR "PAR"OR "PMEPR" OR "peak-to-mean envelope power ratio") AND ("reduction "OR"technique"OR"method"OR"scheme")))AND(english WN LA))
Table 1.3 Summary of search results
Source Extraction date All languages Only English
El Compendex 25-July-2017 511 481
Scopus (All) 25-July-2017 1366 1306
Scopus (TITLE-ALL-KEY) 25-July-2017 303 289
ISI 26-July-2017 204 204
IEEE 25-July-2017 91 90
Total(with Scopus (all)) 2172 2081
Total(with Scopus (TITLE-ALL-KEY)) 1109 1064
1.3.3 Study selection and quality assessment
The Parsifal5, an online tool for systematic literature reviews within the context of Software Engineering, was used to remove duplicates and to execute the systematic literature review.
The study selection was conducted on title and abstracts by applying the following inclusion and exclusion criteria:
• Inclusion criteria:
1. The papers are reported in peer-reviewed workshop conference, journal or technical reports.
2. The papers are written in English.
3. The papers improve the PAPR of the MIMO-OFDM system. However, they do not use a specific PAPR reduction technique.
4. The papers proposed a PAPR reduction solution for a MIMO-OFDM system.
• Exclusion criteria:
1. Full text of studies is not accessible. 2. Studies that are duplicates of other studies. 3. The papers are not described in English.
4. The papers have not been published in a peer-reviewed conference or journal. 5. The papers do not propose a PAPR reduction solution for a MIMO-OFDM system. The number of included, excluded, and duplicated articles is shown in Table 1.4 for each source. It should be noted that apparently, the exclusion list was updated during the full-text reading process for the given criteria.
Table 1.4 Number of studies per study selection
Source Accepted Rejected Duplicated Total
Compendex 136 45 300 481
IEEE Digital Library 10 2 78 90
ISI Web of Science 15 18 171 204
Scopus 56 59 174 289
All Sources 217 124 723 1064
The quality assessment was conducted on the set of selected articles. Four questions were evaluated at this point:
1. Is the research objective a PAPR reduction technique in a MIMO-OFDM system? 2. Is the PAPR reduction technique in the MIMO-OFDM system clearly defined?
3. Does the article compare the performance of the technique presented with the conventional MIMO-OFDM system? (Yes = CCDF and BER, Partial = CCDF or BER, No = none) 4. Has the article been cited (Google scholar)? (Before 2015: Nc = number of citations, if
Nc≥10 then Yes, if 10<Nc≤3 then Partial, if 3<Nc≤0 then No. For 2015 to 2017: if
Nc≤5 then Yes, if 5<Nc≤1 then Partial, ifNc=0 then No)
The possible answers for all quality questions were the same, and are presented with their weights assigned in Table 1.5. The maximum score for the quality assessment was 4, and the cutoff scores, 2. The result after executing the quality assessment for all selected articles is summarized in Table 1.6. Table 1.5 Quality Assessment: Answers Description Weight Yes 1.0 Partial 0.5 No 0.0
Table 1.6 Quality Assessment Score
Quality score Number of articles
4 26 3.5 43 2.5 36 Subtotal(accepted) 163 2 33 1.5 10 0.5 3 0 4 Subtotal(rejected) 54 Total 217 1.3.4 Data extraction
Table 1.7 Data extraction form
Description Kind RQ
1a. Study ID String Field
1b. Article Title String Field
1c. Author name String Field
1d. Year of Publication Integer Field RQ1
1e. Download Database String Field
2a. Kind of MIMO-OFDM system used String Field RQ2
2b. Number of antennas for simulation String Field RQ2
3a. PAPR reduction technique used String Field RQ3
3b. Classification of the PAPR reduction technique:
i Coding (C)
ii Hybrid (H)
iii Multiple Signalling & Probabilistic (MSP)
iv Signal Distortion (SD)
Select One Field RQ3
3c. The PAPR reduction technique is:
i adapted from OFDM
ii takes advantage of the structure of MIMO system
Select One Field RQ4
3d. PAPR reduction technique description String Field RQ3
4a. How much improve the PAPR (CCDF=10−3)? String Field RQ5
4b. Does the technique cause BER degradation?
i No
ii There is no information
iii Yes
Select One Field RQ5
1.3.5 Data analysis
The information collected from the data extraction form for each primary study was tabulated. Tables and illustrations were developed to present the extracted data efficiently (see Section 1.4).
1.4 Results of systematic literature review
1.4.1 Frequency of publication (RQ1)
The analysis of the number of articles per year is shown in Fig. 1.3. We find articles that refer to PAPR reduction techniques in MIMO-OFDM, starting in 2003. The number of articles per
year has increased significantly since 2007, with likely reason being the growing interest in MIMO systems and their combination with OFDM, the inclusion of MIMO-OFDM in all 4th generation (4G) and next generation (5G) wireless communications systems. Note that for the year 2017, the mapping covers only the period before July 25.
2 2 8 8 14 16 14 14 14 11 17 16 8 14 5 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 Num ber of ar ticles Year
Figure 1.3 Final articles per year (2017, before of 25 July)
1.4.2 Download Database
Fig. 1.4 presets the number of articles per download database. It can be seen here that 75% of the articles were downloaded from IEEE explore. On the other hand, the Springer Link, Else-vier, and Wiley online libraries respectively saw a total of 6%, 4%, and 3% of all downloads. All of the articles were from conference papers, journal papers or letters.
1.4.3 MIMO-OFDM system applied (RQ2)
Next, we clarify the taxonomy employed in this work. First, in the MIMO literature, we can see a classification of MIMO systems based on antenna configurations. Four categories are recognized, namely, single-input single-output (SISO), single-input multiple-output (SIMO), multiple-input single-output (MISO), and multiple-input multiple-output (MIMO). However, note that, strictly speaking, a MIMO system should be one with multiples antennas at both the transmitter and receiver (Hampton, 2013).
IEEE explore 75% Elseivier 4% Springer Link 6% Wiley online library 3% Others 12%
Figure 1.4 Articles per download database
On the other hand, MIMO techniques can be divided into two categories according to the tar-get objective, namely, spatial diversity or spatial multiplexing (see Fig. 1.5). Spatial diversity techniques aim to combat fading, and are used to improve reliability and provide diversity gain. With spatial diversity, we send information across the different propagation paths using the space-time coding (STC) schemes. The STC include two major categories: space-time block codes (STBCs) and space-time trellis codes (STTCs). Meanwhile, STBCs are divided into two subclasses, namely, orthogonal space-time block codes (OSTBCs) and non-orthogonal space time block codes (NOSTBCs). Unlike spatial diversity, in spatial multiplexing techniques, we send different portions of information along different propagation paths. Spatial multiplexing is used to increase throughput and provides degrees of freedom (multiplexing gain). There are some schemes that are used to implement spatial multiplexing (see Fig. 1.5), and that are based on layered space-time (LST) coding, such as the Bell Laboratories layered space-time (BLAST) family: horizontal BLAST (H-BLAST), vertical BLAST (V-BLAST), and diagonal BLAST (D-BLAST), multi-group space-time coding (MGSTC), and threaded space-time cod-ing (TSTC) (Hampton, 2013). Additionally, spatial multiplexcod-ing can be implemented uscod-ing Eigen-beamforming (Hampton, 2013).
MIMO OSTBC Quasi-OSTBC STBC STTC H-BLAST V-BLAST D-BLAST MGSTC Spatial Multiplexing Spatial diversity
Figure 1.5 Purpose MIMO taxonomy Hampton (2013)
Another classification is possible, based on the number of users, and in that the MIMO tech-niques can be categorized as single-user MIMO (SU-MIMO) or multi-user MIMO (MU-MIMO). SU-MIMO involves one transmitting node with multiple antennas, and one receiving node (Hampton, 2013). In MU-MIMO, depending on the mobile cellular system, each piece of user equipment (UE) with a single antenna transmits to a Node-B (eNB), and the eNB processes the signals of each UE as if they were coming from multiple transmit antennas (Hampton, 2013). Additionally, we can classify MIMO techniques according to what is know about the transmit-ter and receiver of the channel charactransmit-teristics. In this case, we know two categories: open-loop and closed-loop. MIMO techniques, in which only the receiver requires knowledge of the chan-nel are open-loop. In contrast, in closed-loop techniques, the receiver must send the chanchan-nel information back to the transmitter.
Finally, we introduce the term massive MIMO, which refers to a MIMO system with a large number of antennas, which can go into the hundreds or thousands, for example, and that can multiply the system capacity, and improve the performance of the communication system in terms of data rate and reliability.
The MIMO-OFDM systems described in the reviewed articles were grouped according to the scheme used, and the result is presented in Fig. 1.6. It can be seen that the PAPR techniques providing the greatest amount of reductions are implemented in a generic model (61 articles) or
STBC MIMO system (60 articles). However, PAPR reduction techniques implemented in MU-MIMO, Massive MIMO or V-BLAST are also available. Note that in some cases, the reviewed studies may apply to one or more types of MIMO schemes. Here, when the study is applied to more than one type of MIMO system, we consider the one with the greatest emphasis in the work. 2 2 61 19 8 6 61 4 V-BLAST STFBC STBC SFBC Multi-user (MU) Massive MIMO Generic model Beamforming (BF) 0 10 20 30 40 50 60 70 MI MO sy st em sc hem e Number of articles
Figure 1.6 Articles per MIMO system type
1.4.4 Proposed classification of PAPR reduction techniques (RQ3)
The techniques proposed in the selected papers were classified based on the taxonomy pre-sented by Sandoval et al.(2017), where the PAPR reduction techniques are divided into four categories, namely, Coding (C), Multiple Signaling & Probabilistic (MSP), Signal Distortion (SD), and Hybrid (H).
Fig. 1.7 shows that 83% of the techniques proposed in the selected articles can be considered as Multiple Signaling and Probability methods, 9% as Hybrid, 6% as Signal Distortion, and only 2% were Coding techniques. Meanwhile, Table I-1 presents primary studies grouped by PAPR reduction method and category. For instance, in the Multiple Signal and Probabilis-tic categories, the most important schemes found are: Active Constellation Extension (ACE),
Cross-Antenna Rotation and Inversion (CARI), Cross-Frequency Permutation and Inversion (CFPI), Precoding, Partial Transmit Sequence (PTS), Selected Mapping (SLM), Tone Reser-vation (TR), and Wavelet Entropy Algorithm (WEA). On the other hand, Signal Distortion techniques as clipping and peak cancellation methods applied to MIMO-OFDM systems were found in the literature. Additionally, a bar graph including the number of first studies per PAPR reduction technique is shown in Fig. 1.8. For instance, 37% of articles analyzed correspond to the SLM and PTS techniques. Twenty articles are categorized as precoding schemes, and fourteen as hybrid techniques. More details regarding the resulting PAPR reduction techniques are presented in the following section.
Coding 2% Hybrid9% Multiple Signalling & Probabilistic (MSP) 83% Signal distortion 6%
Figure 1.7 Articles per proposed PAPR reduction technique type
1.4.5 Classification of MIMO-OFDM PAPR reduction techniques based on its
adapta-tion to the MIMO structure (RQ4)
Numerous studies have attempted to explain the PAPR problem in a MIMO-OFDM system. They investigate PAPR reduction techniques for a MIMO-OFDM system and propose the ex-tension of OFDM PAPR reduction techniques as PTS, SLM, or TR.
2 6 6 30 25 20 3 36 14 4 7 6 4 0 5 10 15 20 25 30 35 40
Wavelet entropy algorithm Tone reservation SLM and PTS SLM PTS Precoding Peak cancellation Others Hybrid Coding Clipping CARI or CFPI Active constellation extension
Number of articles P A P R Reduction T ech ni qu e
Figure 1.8 Articles per PAPR reduction technique
In recent years, a few authors began to propose alternatives for PAPR reduction that tried to exploit the potential of MIMO systems. Khademi et al.(2012) introduced the precoding technique that exploits the Eigen-beamforming mode (EM) in MIMO systems. Using beam-forming can significantly improve the received SNR of OFDM systems, and it has been widely adopted in modern MIMO-OFDM. However, the beamforming deteriorates the PAPR because after beamforming the dynamic range of the signals increases (Hung & Tsai, 2014). For this reason, Hung & Tsai (2014) proposed a new algorithm for single-user MIMO-OFDM systems when using beamforming schemes, i.e., maximum ratio transmission (MRT) and equal gain transmission (EGT), which try to adjust the power at some subcarriers after beamforming. The results of Hung & Tsai (2014) show PAPR reduction, and in addition, improve the bit error rate performance. On the other hand, Prabhu et al. (2014) analyzed the PAPR problem on massive MIMO-OFDM, and introduced a low-complex PAPR scheme, in which a combination clipping and antenna reservation approach is used to reduce the PAPR. Pandurangan & Peru-mal (2011) introduced a modified PTS with forward error correcting codes for PAPR reduction in a MIMO-OFDM system and showed that unlike the original PTS technique, the combina-tion of a modified PTS and FEC provides better PAPR reduccombina-tion and moderate computacombina-tional complexity in MIMO-OFDM systems.
In the systematic literature review, the articles were classified based on whether the technique proposes an extension of OFDM PAPR reduction techniques or if the scheme exploits the MIMO potential. The result is presented in Fig. 1.9, where is can be seen that the current trend is to take advantage of the MIMO structure in designing the adopted PAPR reduction technique. adapted from OFDM 39% takes advantage of the structure of MIMO system 61%
Figure 1.9 Articles per technique approach
1.4.6 Most used PAPR reduction techniques (RQ3) description
1.4.6.1 Coding techniques
The forward error correction (FEC) block can be used to reduce the PAPR if we select the codewords that minimize the PAPR for transmission (Sandoval et al., 2017). These are four articles that describe PAPR reduction schemes based on coding: these include turbo codes (Al-akaidi et al., 2006), convolutional codes (Venkataraman et al., 2006), Reed-Solomon codes (Fischer & Siegl, 2009), Reed-Muller (RM) codes and complementary sequence codes (Jin-long & Yuehong, 2009).
